Maternity and Ethnicity in Scotland

Chalmers J, Bansal N, Fischbacher CM, Steiner M, Bhopal R, on behalf of the Scottish Health and

Ethnicity Linkage Study

Why we should have ethnicity data

• Measure health inequalities and inequity • Monitor impact to reduce these inequalities• Respond to legislation and policies on human

rights and equality in health and health care• Develop and test scientific hypotheses on

disease and risk factor variations

Anonymised Linkage of Health Databases to Census Databases:

conceptualising the procedure

Health Database Census Database

Record Linkage

Encrypted CHI Number

Personal Identifiers

Personal Identifiers

EncryptedCensusNumber

Encrypted CHI Number Encrypted Census Number

(Look-up Table)

Death & Hospitalisation from Health databases

Ethnicity from Census

Background

• Poor recording of ethnicity related to births in Scotland• How do the different ethnic groups in Scotland compare

for various measures of maternity activity• Do the ethnic minority groups in Scotland behave

similarly to the same groups in England– England has poorer maternity data but better ethnicity

data• Millennium cohort• NHS Numbers for Babies (NN4B)

Methods

• Population – women in Scotland on census night who subsequently deliver a baby (up to April 2008) in Scotland recorded on SMR02 (98% of all deliveries)

• Restriction to first baby– “Purer” data

• Subsequent deliveries strongly influenced by first delivery

– Particularly caesarean sections

• Parous women different from nulliparous

Overall results

• 95% of 2001 population of 4.9 million linked• 363,990 records from new SMR02 file that fall

in study period (May 2001 to April 2008) – Of these 192,803 link to census and are first

single birth records only.

Areas of interest

• Maternal age– Strong influence on risk of caesarean

section• Smoking

– Strong influence on birthweight• Type of delivery• Analgesia • Birthweight and Gestation

– Previous evidence of ethnic differences

Numbers

Ethnic Group 

White Scottish 170803

Other White British 12992

White Irish 1681

Other White 3159

Indian 537

Pakistani 1689

Other South Asian 295

Chinese 483

Black 307

Any Mixed Background 440

Other Ethnic Group 416

First Births - mean age (with 95% Cis)

24

25

26

27

28

29

30

31

32

33

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Maternal age at first birth

• Note general association between deprivation and low maternal age– May explain older “white immigrants”

• Pakistani women tend to be young– Also noted in millennium cohort

Percentage of women smoking during pregnancy

0

5

10

15

20

25

30

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Smoking in pregnancy

• Higher in White groups• Similar pattern to studies in England

Percentage using strong or mild analgesia in labour

0

10

20

30

40

50

60

70

80

90

100

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Strong

Mild

Analgesia in labour

• No particular pattern

Caesarean section rates

• No statistically significant difference– Contrasts with higher rates noted in most of

the ethnic minority groups in a London cohort (Ibison 2005)

• Perhaps White rate has caught up

Birthweight and gestation

• Complex area – mechanisms poorly understood• Gestation has major effect on birthweight but not all

differences in birthweight explained by gestation– Physiologically

• Balance between baby’s “desire for optimal environment” and mother’s “accommodation” and ability to deliver safely

– Epidemiologically• Genetic factors (influencing maternal and baby size and birth

mechanisms), nutrition, smoking, specific illnesses, deprivation, intervention etc

• Important to discern effect of ethnicity

Odds ratio of preterm delivery, adjusted for age, deprivation and smoking

0

0.5

1

1.5

2

2.5

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Gestation

• Pakistani babies have shorter gestation• Observed in millennium cohort, along with most

minority ethnic groups• Interest in generational effect

Mean Birthweight (grams) adjusted for maternal age, with 95% CIs

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Effects of various adjustments on birthweights (grams)

2900

3000

3100

3200

3300

3400

3500

3600

White

Sco

ttish

Other

White

Brit

ish

White

Iris

h

Other

White

India

n

Pakis

tani

Other

South

Asi

an

Chines

e

Black

Any M

ixed

Bac

kgro

und

Other

Eth

nic G

roup

Unadjusted

Adjusted for maternal and gestationalage

Adjusted for maternal and gestationalage and smoking

Adjusted for maternal and gestationalage, smoking & deprivation (SIMD)

Birthweight – multivariate analysis

• Pakistani and Indian – 260 grams lower• Black – 119 grams lower• Maternal age – 5.8 grams increase with each year• Gestation – 183 grams increase with each week• Smoking – 275 grams reduction• Deprivation – least deprived decile 185 grams heavier

than most deprived decile• No “migrant status” effect

Birthweight

• All non-white groups seem to have smaller babies– Some of this effect mediated by shorter gestations– Minimal influence of smoking and deprivation– Effect seen even when restricted to term babies– Difficult to adjust for interventions

• General agreement with other UK data

Conclusions

• Some ethnic variations in maternity measures• Generally aligned with results from other UK

studies• Specific effects on birthweight and gestation• Important for predicting prenatal growth.